**** BEGIN LOGGING AT Thu Jul 24 03:00:02 2014 Jul 24 06:49:37 Hi yocto.. Jul 24 06:55:58 Hi yocto...I am trying to interface camera with beaglebone with support of mplayer but i am getting paste bin error how to rectify this error http://pastebin.com/KigCzmmW any one can help to resolve this. Jul 24 07:17:48 good morning Jul 24 07:27:37 Hi everyone Jul 24 07:31:05 I'm getting this error when I build a custom image: Jul 24 07:31:07 arm-poky-linux-gnueabi-ld: unrecognized option '-Wl,-O1' Jul 24 07:31:22 anybody have an idea about where that originates? Jul 24 07:38:07 Hi please any one help me to solve this camera issue. Jul 24 07:49:26 andyj25: which recipe is failing? What was it building? This is an option for CC (gcc) to be passed to the linker LD Jul 24 07:51:44 Hi ant_work. I made my own layer, and a recipe for a very simple image. Jul 24 07:52:19 When i do bitbake myimage, it fails during do_compile() saying "unrecognized option '-Wl,-O1'" Jul 24 07:52:26 is it something to do with LDFLAGS? Jul 24 07:52:33 yes Jul 24 07:53:40 I think the "-Wl, -O1" options come from poky/meta/conf/bitbake.conf Jul 24 07:53:52 should i be trying to overwrite that from my image recipe file? Jul 24 07:53:55 gcc -Wl,foo,bar -> ld foo bar Jul 24 07:54:55 you must have changed LDFLAGS or EXTRA_OEMAKE Jul 24 07:55:00 in one recipe Jul 24 07:55:29 which one is failing? the image itself does not compile anything Jul 24 07:57:09 the image recipe builds a bunch of libraries (which succeed) then it tries to build another recipe which I wrote Jul 24 07:57:17 that recipe is PJSIP Jul 24 07:57:24 and that's where it fails compiling Jul 24 07:59:23 I haven't written an LDFLAGS or EXTRA_OEMAKE in that recipe file Jul 24 07:59:58 please check the build logs and verify CFLAGS and LDFLAGS Jul 24 08:00:12 of that recipe Jul 24 08:00:25 OK. one moment. Jul 24 08:00:45 temp/log.do_compile.xxxx Jul 24 08:00:46 that one right? Jul 24 08:03:44 my temp/run.do_compile says ---> export TARGET_LDFLAGS="-Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed" Jul 24 08:04:20 and also ---> export LDFLAGS="-Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed" Jul 24 08:06:44 what is LD? Jul 24 08:07:21 this would only work if you use gcc for linking Jul 24 08:08:21 I suspect you should verify the Makefile Jul 24 08:10:51 andyj25: I've never built this source. Seems it was an oe-classic recipe Jul 24 08:10:56 http://tinyurl.com/qc3x2zd Jul 24 08:12:05 LD is this ---> LD="arm-poky-linux-gnueabi-ld --sysroot=/home/andrew/yocto-imx6/build/tmp/sysroots/imx6qsabresd" Jul 24 08:12:44 so the Makefile for the source code I'm compiling can also modify LDFLAGS? Jul 24 08:17:39 Thanks for the pointer to that sflphone recipe.. i'm looking into it for clues Jul 24 08:19:15 andyj25: sorry, phone call Jul 24 08:19:46 you should normally use the default flags passed by OE/Yocto toolchain Jul 24 08:20:55 some sourcess though need to be patched or you can adjust in the recipe i.e. with EXTRA_OEMAKE Jul 24 08:21:02 could it be possible that the default flags don't work with that cross-compiler? Jul 24 08:21:22 or do you think the defaults are being overwritten by something Jul 24 08:21:28 please check the logs of another recipe, you'll see Jul 24 08:28:22 alright. thanks for the tips. Jul 24 08:28:47 so you mean.. i should compile another recipe which works Jul 24 08:28:49 and then check its logs? Jul 24 08:33:05 yes, as general case every recipe should use the pre-defined ${CC} ${CFLAGS} ${LDFLAGS} Jul 24 08:36:04 morning all Jul 24 08:36:27 hi bluelightning, all Jul 24 08:36:33 hi mckoan Jul 24 08:36:40 gm bluelightning Jul 24 08:37:13 hi ant_work Jul 24 08:41:23 andyj25: the story is a bit long, per default we pass '-e' to make Jul 24 08:41:55 "Give variables taken from the environment precedence over variables from makefiles" Jul 24 08:42:31 you could try redefining the full EXTRA_OEMAKE Jul 24 08:43:00 ohh.. so Yocto passes "-e" by default? Jul 24 08:43:22 yes, check bitbake.conf Jul 24 08:44:32 but there are exceptions, I can't find a better documents offhand Jul 24 08:44:34 http://lists.openembedded.org/pipermail/openembedded-devel/2011-May/078184.html Jul 24 08:44:42 check this old thread pls Jul 24 08:52:43 ok thanks : ) Jul 24 08:53:32 Hi please any one help me to solve this camera issue. Jul 24 08:54:30 Hi yocto...I am trying to interface camera with beaglebone with support of mplayer but i am getting paste bin error how to rectify this error http://pastebin.com/KigCzmmW any one can help to resolve this. Jul 24 08:54:33 Hi please any one help me to solve this camera issue. Jul 24 09:32:37 Hi all =) Jul 24 09:41:18 ant_work, I found a thread from the mailing list which describes exactly the problem i'm getting Jul 24 09:41:21 https://lists.yoctoproject.org/pipermail/yocto/2013-January/011581.html Jul 24 09:42:24 According to kernel.bbclass, setting EXTRA_OEMAKE = "" is one way to get rid of the "-e" option, but that didn't work for me Jul 24 09:57:45 andyj25: is that source? https://github.com/chakrit/pjsip Jul 24 09:59:54 that link is a fork Jul 24 10:00:02 i'm using the official source code Jul 24 10:00:19 ok, it is an auto-tooled recipe Jul 24 10:00:19 but anyway, it looks similar to that Jul 24 10:00:41 this implies EXTRA_OEMAKE = "" Jul 24 10:00:47 but that's not th epoint here Jul 24 10:01:28 the recipe offers an hook, see user.mak.sample Jul 24 10:07:31 you can probably patch that Jul 24 10:07:43 you have to force $(CC) $(LDFLAGS) Jul 24 10:30:39 hmm.. sorry but i don't quite understand Jul 24 10:30:44 does that mean i should make a user.mak file? Jul 24 10:30:50 in the pjsip source? Jul 24 11:09:58 hey guys Jul 24 11:11:17 I've gtot another question. Should be really easy for you - if I want to clean up be deploay dir completely without touching any sources or fetched/compiled data while the build process, which bitbake -c clean command can I use to get all of it together Jul 24 11:13:01 by now I just removed my /tmp dir to get back to a clean environment Jul 24 11:13:19 is it better to use a bitbake cmd ? Jul 24 11:17:04 depends Jul 24 11:17:16 what's your reason for wanting to do the cleaning? Jul 24 11:27:05 just to delete all images which just were tests while the developing process, I allways add a package or a recipe and test it if it works I go on. So just not to mix images up because de date code is no really very clearly represented Jul 24 11:27:18 no special reason just for the overview =) Jul 24 11:35:49 JaMa: is that a record low for failed builds in world? Jul 24 11:41:24 andyj25: sorry but I cannot check the sources right now (nor the recipe you're building ;). You have to investigate the direct invocation to LD. Possible test: 1) set CFLAGS += LDFLAGS and LDFLAGS="" in the recipe. This way LD should not protest. 2) Fix the sources, probably one Makefile 3) patch the autoconfig files to add LDFLAGS to CFLAGS like 1) Jul 24 11:42:18 andyj25: there are som eold recipes for oe-classic around, check how they did Jul 24 11:44:03 ok Jul 24 11:44:12 thanks for all the tips! Jul 24 11:45:20 try patching the user.mak in that way at first Jul 24 11:45:29 and verify the recipe inherits autotools Jul 24 11:46:24 btw, this seems rather complicated recipe, with different flags for lib and apps Jul 24 11:46:36 it is not hello-world ;)ù Jul 24 12:03:12 motzer_: generally you can just delete the old images, there isn't a command to delete them on demand Jul 24 12:03:33 motzer_: however you can set RM_OLD_IMAGE = "1" in which case old images will be deleted automatically each time a new one is created Jul 24 12:05:08 RP: no, 2014-05-15 was better :) Jul 24 12:09:26 JaMa: ah, true :). Its getting close again at least Jul 24 12:09:28 RP: and abiword + libssh were blacklisted because of libgrcypt upgrade since then http://www.openembedded.org/wiki/Bitbake_World_Status#Failed_tasks_2014-05-15 Jul 24 12:09:54 JaMa: hi. Unfair of you doubling the klibc-utils warning checking both 2.0.3 and 2.0.4 :p Jul 24 12:09:56 RP: thanks for all the patches you've sent Jul 24 12:10:15 JaMa: np, its good to see its helped a bit :) Jul 24 12:10:23 ant_work: you'll get double kudos if you fix it :) Jul 24 12:10:31 JaMa: floating dependencies next I guess... Jul 24 12:10:48 * RP keeps fixing the oe-core ones he sees in local builds Jul 24 12:11:05 RP: I've included all recipes reported by insane_qa to next test-dependencies build to see why we're seeing different results from these 2 checks Jul 24 12:11:21 RP: I'm scared by that /bin/bash issue revealed by Robert's testing Jul 24 12:11:51 ant_work: Sadly, I'm not, I've known we've had some issues there for a while. Its why bugs were opened and people are looking at it Jul 24 12:11:52 he is using bash on his Ubuntu 12.04 afais Jul 24 12:12:18 RP: the sad part is that we're both quite busy with other patches, it would be great to get more fixes from people actually using these recipes on target Jul 24 12:12:35 RP: xz uses a posix-shell.m4 Jul 24 12:12:56 JaMa: yes, that is sad and frustrating Jul 24 12:13:32 ant_work: I know I had to do some nasty things with libtool to resolve the problems there Jul 24 12:13:58 and I'm still a bit concerned about that foreign patch introducing new COPYING files with GPLv3 :/ Jul 24 12:14:09 (it coded into the script whether /bin/sh handled bashisms or not) Jul 24 12:14:36 I've read about a checkbashism tool Jul 24 12:14:53 but I've finally fixed all our internal components to at least not fail in do_configure Jul 24 12:14:55 maybe we could try to sanitize *some* scripts Jul 24 12:15:16 JaMa: in some ways it is better to expose this problem and deal with it rather than hiding it though... Jul 24 12:15:35 JaMa: you can't guarantee software would always be built by our hacked automake Jul 24 12:15:39 yes, but sometimes it's exposed in "hidden" way :) Jul 24 12:15:55 JaMa: its far from perfect, I agree Jul 24 12:18:07 JaMa: bluelightning: pls push klibc 2.0.4 so I can send the fix on top of it Jul 24 12:19:12 ant_work: rebase your change on top of master-next Jul 24 12:19:20 mabe later today I'll send the quick fix then I'll see if a refactoring is possible Jul 24 12:19:39 ant_work: I plan to push it after next build, if you send patch on top of it I'll include it as well Jul 24 12:19:49 the split klibc / klibc-utils is OE specific. It is a single package in debian Jul 24 12:20:13 (we add klcc-cross and the -static utils) Jul 24 12:20:18 I have been bitbakin qt 5.2.1 in my build without any issue until I added qtquickcontrols-qmlplugins. Now I get this error: * opkg_install_cmd: Cannot install package qtquickcontrols-qmlplugins. Is there a known issue with this? Can someone point me to a patch I can apply? Thanks! Jul 24 12:20:41 JaMa: ok Jul 24 12:21:20 TRoGd0R: check which packages were created by qtquickcontrols Jul 24 12:21:29 How do I do that? Jul 24 12:21:36 TRoGd0R: it could be empty (not created) with your qtbase's PACKAGECONFIG Jul 24 12:21:42 Sorry I am new to this Jul 24 12:22:35 easiest would be find tmp-eglibc/deploy/ipk -name \*quickc\* Jul 24 12:23:15 more correct would be to check qtquickcontrols WORKDIR packages-split directory Jul 24 12:23:20 RP: about the tar packages, note there are layers carrying around "deltask do_package_write_tar" Jul 24 12:23:40 (maybe just meta-initramfs hopefully) Jul 24 12:25:52 ant_work: it shouldn't cause any harm even if we do delete it Jul 24 12:26:31 and I was just hit by the changes to insane.bbclass adding do_package_qa... ;) Jul 24 12:27:22 JaMa: I'll send the possible fix for the kexecboot kernel together ^^ Jul 24 12:27:26 ant_work: the joys of maintainership! ;-) Jul 24 12:27:40 ant_work: the improvements to dependency validation should make that worthwhile Jul 24 12:27:50 it would be easy if it were possible to grep on *all* layers ;) Jul 24 12:29:22 ant_work: if you want to add some kind of full-text-indexing + search to the layer index, please do ;) Jul 24 12:30:49 I can imagine it's mysql structure ;) Jul 24 12:32:08 the app uses django, but you'd have to add some kind of full-text-indexing engine to it Jul 24 12:32:22 of which there are many off-the-shelf solutions Jul 24 12:32:46 JaMa: http://pastebin.com/79kLZ51Z Jul 24 12:33:08 Hi yocto...I am trying to interface camera with beaglebone with support of mplayer but i am getting paste bin error how to rectify this error http://pastebin.com/KigCzmmW any one can help to resolve this. Jul 24 12:33:10 Hi please any one help me to solve this camera issue. Jul 24 12:33:25 It took awhile becuase I searched the entire build directory, I didn't have a tmp-eglibc directory Jul 24 12:34:27 embed: this isn't really the forum for such problems I'm afraid; you should try asking somewhere related to the software you are having problems with (mplayer?) Jul 24 12:35:16 bluelightning: I mean you should have every file stored as text in a field, then you can perform string search on it Jul 24 12:35:59 bluelightning: I guess layers-index just keeps th elinks Jul 24 12:36:01 ant_work: that's pretty slow to search though, you really need a proper index Jul 24 12:36:05 ant_work: that's correct Jul 24 12:36:52 bluelightning: there are powerful functions Jul 24 12:45:36 JaMa: more about /linux-yocto-tiny-kexecboot_3.10.bb, do_package_qa Jul 24 12:46:24 the exposed task reveals that qemux86 is trying to use 3.14 (preferred) kernel while building a 3.10 recipe Jul 24 12:46:59 It could be interesting to understand why before silencing the QA Jul 24 13:09:54 TRoGd0R: you probably don't have icu enabled in qtbase Jul 24 13:10:50 I do I enable that? Jul 24 13:11:35 All I did was to add qtquickcontrols-qmlplugins to my IMAGE_INSTALL_append in local.conf Jul 24 13:11:44 I didn't know there was another step :) Jul 24 13:13:58 So I just went into my qtbase_5.2.1.bbappend file and I see that I have this line DEPENDS += "icu libgles-omap3" Jul 24 13:15:30 that isn't correct, icu should be enabled in PACKAGECONFIG Jul 24 13:15:36 read meta-qt5 wiki and qtbase.inc Jul 24 13:22:51 Hi, does anybody know how to make my mouse work in yocto? Jul 24 13:22:52 [ 1027.105] (WW) Warning, couldn't open module evdev Jul 24 13:23:01 [ 1027.105] (II) UnloadModule: "evdev" Jul 24 13:23:01 [ 1027.105] (II) Unloading evdev Jul 24 13:23:01 [ 1027.105] (EE) Failed to load module "evdev" (module does not exist, 0) Jul 24 13:23:01 [ 1027.105] (EE) No input driver matching `evdev' Jul 24 13:23:08 the architecture is armhf Jul 24 13:23:26 I see there is a line like this in qtbase.inc PACKAGECONFIG[icu] = "-icu,-no-icu,icu" and there is a comment about how qtwebkit needs icu but I am not seeing how I should do anything different than what I am doing. Sorry for my n00bishness Jul 24 13:25:47 TRoGd0R: you need PACKAGECONFIG_append = " icu" in your .bbappen Jul 24 13:25:52 .bbappend Jul 24 13:25:54 LocutusOfBorg1: install xf86-input-evdev Jul 24 13:26:20 OK cool let me try that, thanks! Jul 24 13:34:49 thanks jackmitchell Jul 24 13:34:53 thanks JaMa Jul 24 14:26:16 I’m preparing Yocto build for a custom AM335x Texas Instruments board, what do I have to do with mux.h and pinmux.h from the PinMux utility? Jul 24 14:26:28 I changed mux.h in kernel’s arch/arm/mach-omap/ Jul 24 14:26:42 But i don’t understand how/where to include the pinmux.h Jul 24 14:26:46 any suggestion? Jul 24 14:32:22 bluelightning: thanks good information =) Jul 24 14:33:58 I updated my qtbase_5.2.1.bbappend and removed icu from DEPENDS and added it to PACKAGECONFIG_append. I am still getting the same error, here is the build output, any other ideas? http://pastebin.com/2P6iLNe3 Jul 24 14:47:52 zeddii: so linux-yocto-dev is at 3.16-rc6 Jul 24 14:49:49 TRoGd0R, make sure you have PACKAGECONFIG[accessibility] enabled Jul 24 14:50:59 ant_work. wow. I pushed that about 5 mins ago! Jul 24 14:51:02 Hey all. Having trouble locating the problem: kernel modules aren't getting stripped. (Dora) Jul 24 14:51:26 I'm testing yaffs2 now. I didn't have it in my first build cycles, I spent a lot of time fighting with lttng. Jul 24 14:51:32 * zeddii finishes up his pull request Jul 24 14:53:51 kroon: How do I do that? Jul 24 14:55:17 TRoGd0R, PACKAGECONFIG_append = "accessibility" in the .bbappend. Or you can do PACKAGECONFIG_append_pn-qtbase = "accessibility" in your distro.conf Jul 24 14:56:10 _append = " accessibility", with the prepended blank space, just to be sure.. Jul 24 14:58:44 Cool I just updated to this PACKAGECONFIG_append = " icu accessibility" Jul 24 14:58:56 should that work? Jul 24 14:59:18 zeddii: some patches have been hit by /drivers/video -> drivers/video/fbdev Jul 24 14:59:24 TRoGd0R: yes Jul 24 14:59:44 ok it's building now Jul 24 14:59:55 TRoGd0R, icu is in there by default, but otherwise yes Jul 24 15:01:18 oh, its not Jul 24 15:02:10 hello, is Phil Blundell lurking here ? Jul 24 15:03:44 http://cgit.openembedded.org/openembedded-core/commit/?id=613d0275a59be9154c2a2cdc101bd3e2c79a6d44 -> I wonder why the sticky bit was deemed useful in busybox a bit more than four years ago? Jul 24 15:04:00 (isn't it a kind core security hole?) Jul 24 15:04:05 kinda* Jul 24 15:23:51 RP: I'm sorry to see that the NEEDINFO task about multi-machines builds fall again on you, but on the other hand I'm glad that you'll bring other view on this (so that official Yocto documentation doesn't just copy how I do multi-machine build) Jul 24 15:38:21 JaMa: Its a case of "someone" needing to put info together and you/I appear to be the only people who can :/ Jul 24 16:56:21 can someone check if their /build/downloads/ folder also contains a file called "*.done" Jul 24 16:56:31 (ls /build/downloads/\*.done Jul 24 16:57:06 not here Jul 24 16:57:28 nor here Jul 24 16:57:48 strange, all my downloadfolders have it Jul 24 16:58:17 maybe because I use SRC_URI = "file://*" a couple of times Jul 24 17:00:39 and rsync seems to have problems with it when trying to move the download folders to the download mirror Jul 24 17:51:30 is there a way on the generated target image after flashing to figure out whether a binary is stripped? Busybox does not seem to have the "file" command available. Jul 24 17:53:04 use binutils to determine if the binary has debug sections? Jul 24 17:53:08 or look at the size Jul 24 17:53:13 or install file Jul 24 17:53:25 track down the binary on your build system and look at it there Jul 24 17:53:25 should be a function of config settings/packaging in the build... Jul 24 17:53:26 install file is not an option without development environment Jul 24 17:53:29 size does not tell me much Jul 24 17:53:34 are you seeing something unexpected? Jul 24 17:53:55 not sure what binutils means in this context. Jul 24 17:54:12 nm or readelf maybe? Jul 24 17:54:22 er, scanelf? Jul 24 17:56:39 those are not available, at least on the core-image-minimal Jul 24 17:56:54 lpapp, you will not be able to do this from within the image you create Jul 24 17:57:04 no access to package feed? Jul 24 17:57:04 you need to look at the binaries on the machine you used to create the image Jul 24 17:57:05 ok, thanks (sad). Jul 24 17:57:07 does that make sense? Jul 24 17:57:29 well, sorta. I will file a feature request against busybox to create a file applet ;-) Jul 24 17:57:35 or manually copy the binutils package and install it... Jul 24 17:58:07 how did you get this core-image-minimal anyway? did you build it yourself or download it from somewhere Jul 24 18:01:32 I dont understand how to make a simple gui program in yoctoproject. I have looked at lots of docs and it seems heavily focused on the architecture of a program. Jul 24 18:02:00 but I havent been able to find any resources that completly detail creating a custom application Jul 24 18:02:14 do you have an existing GUI program and you want to make a recipe for it so you can put it on images? Jul 24 18:02:24 or are you trying to make a GUI front end for yocto? Jul 24 18:02:43 (really Im interested in a simple charting engine (that streams data in) (from a usb connector)) Jul 24 18:03:20 no Im trying to create basically a hello world type application just to learn about it Jul 24 18:04:07 so yocto is not really a way to create new programs, it's a way to put lots of existing programs together Jul 24 18:04:11 Im really familliar with python , fairly proficient with c/c++ , marginally proficient with Java ... but I just dont really understand how to make a simple yoctoproject gui Jul 24 18:05:01 by programs you mean "recipes"? Jul 24 18:05:39 or actual programs? that the recipes tell to execute? Jul 24 18:05:46 yeah, actual programs Jul 24 18:06:11 maxtothemax: I am just the user of the image. I honestly think busybox ougtha have the "file" applet, anyhow. Jul 24 18:06:54 so Ive found plenty of tutorials that show writing a simple "helloworld.cpp" and then executing it (since its essntially a stripped down linux from what I can tell) Jul 24 18:07:25 lpapp, I doubt they'll put one in, but what you can do is boot up your image, copy the file you're interested in onto a flash drive or do a network copy or something, and then look at it on a more full-featured Linux OS Jul 24 18:07:57 maxtothemax: I doubt they would not put one in, especially configurably. Jul 24 18:08:08 lpapp, or if you have a copy of your image on a Linux machine, you can mount it with loopback and look at the files that way Jul 24 18:09:05 joranb, yocto is designed to create a complete Linux OS with a custom loadout of preinstalled programs on it Jul 24 18:09:37 joranb, you know how some routers have Linux operating systems on them? Jul 24 18:09:58 a lot of those operating systems were generated using Yocto Jul 24 18:10:06 yes Jul 24 18:10:39 and when a firmware update for that router comes out, the replacement operating system was also generated with yocto Jul 24 18:10:43 that's sort of what it's for Jul 24 18:11:09 ok ... so gumstix pepper board is what Im developing for ... I would likely just put my program execution into the ~/.bashrc or something Jul 24 18:11:24 joranb, now we're talking :) Jul 24 18:11:38 here's what I would suggest Jul 24 18:11:56 install a yocto image on your gumstix Jul 24 18:12:04 but I dont understand what graphics libraries I can use (ie wx,qt,etc) Jul 24 18:12:20 uh huh I can manage that much Im sure ;P Jul 24 18:12:49 then develop your program as you normally would, using your PC for testing of your program Jul 24 18:13:11 finally, write a yocto recipe for your program Jul 24 18:13:18 and generate a new yocto image that includes your program Jul 24 18:13:28 as for graphics libraries Jul 24 18:13:33 I know we have recipes for GTK Jul 24 18:13:39 ahhh Jul 24 18:13:40 perfect Jul 24 18:13:49 we might also have recipes for WX and QT Jul 24 18:13:54 that makes sense then :) Jul 24 18:14:17 ok I think I get it now ... maybe ... Jul 24 18:14:34 I guess ill try this weekend and see how far I get Jul 24 18:14:43 oe-core is all the recipes that are included with yocto: https://github.com/openembedded/oe-core Jul 24 18:14:50 I'll likely be back with more specific questions this weekend Jul 24 18:14:57 ahh that helps alot I think Jul 24 18:14:59 but there are lot's more Jul 24 18:15:10 look in metal-oe, it's a huge library of extra recipes: https://github.com/openembedded/meta-oe Jul 24 18:15:44 you can add the meta-oe layer to your bblayers.conf, and you'll be able to use whatever of those extra recipes you want Jul 24 18:15:48 I will definately do that Jul 24 18:15:52 It makes more sense now Jul 24 18:15:57 that you explained some stuff Jul 24 18:16:12 I was just more confused by than anything by the getting started guides Jul 24 18:16:36 yocto is kind of unique, it took me a couple of days before I understood what it was and what it was for Jul 24 18:18:08 I think I get it now ... my program is just to stream values from a usb tty and plot them (animated) and show the latest value at the top ... and thats all its supposed to do (really thats all the whole device is supposed to do) Jul 24 18:18:31 which is really simple ... but the yocto ecosystem is the hardest part ... Jul 24 18:18:32 and if you want to generate a little mini-OS that will start up your program at boot time and do nothing else, yocto is perfect Jul 24 18:19:11 thats exactly what I want to do ... I have done it in android on touchrevolution , but I was interested to try out yocto and gumstix Jul 24 18:19:28 assuming you have a PC with Linux on it, you should develop your program on that and then make a yocto recipe for it to get it on your gumstix Jul 24 18:19:47 got It ... I think Jul 24 18:20:11 I will make a super simple ui this weekend and let you know ... or come back with more specific how do i type questions :P Jul 24 18:20:11 maxtothemax: it kind of depends on what your goals are. yocto is actually better for maintaining your image for your device long term. if all you're doing is something extremely light, e.g. board bringup or just one thing to run on the target, buildroot is faster and lighter. but yocto makes it easy to maintain your image and packages for the long term, and to maintain a linux distribution for embedded Jul 24 18:22:49 Thanks alot for the help ... I think I understand better what yocto is now :) Jul 24 18:23:53 @kergoth never heard of buildroot , just googled it ... I'll have to look into it further Jul 24 18:24:05 how hard is it to build an image for gumstix with it? Jul 24 18:25:19 there's no guarantee it supports gumstix out of the box at all. nearly every device has its own kernel and patches and configuration. any tool targeting embedded devices needs explicit support for whatever it is, generally Jul 24 18:29:00 meh ok ... well I know I can get a gumstix layer for yocto definately (so I guess I'll stick with that for now... maybe when I have a better handle on these builds I'll revisit buildroot) Jul 24 18:29:21 thanks though Jul 24 18:29:33 I'll file that bookmark away for later examination Jul 24 18:44:02 rburton: did you ever look back at the automatic .debug for FILES_${PN}-dbg bits that you hacked together? Jul 24 18:46:01 Anyone have thoughts on why my kernel modules are not being stripped? Jul 24 18:48:08 ah, there's the contrib branch Jul 24 18:50:30 WarheadsSE, none of my kernel modules in FC20 are stripped Jul 24 18:50:51 WarheadsSE, are kernel modules allowed to be ? Jul 24 18:51:38 they can be stripped, but it has to be done more carefully than the strip of regular binaries Jul 24 18:52:06 afaik it should be done automatically int he build process, not sure why yours wouldn't be Jul 24 18:53:08 by "not stripped", I mean at least that what "file" says :) Jul 24 18:53:43 you can't strip all symbols, they're needed to load the module. but iirc debug info and whatnot can be Jul 24 18:56:44 kergoth: sadly not Jul 24 19:00:08 would love to see that branch get merged :) definitely nicer than hardcoding the list, given it already knows about the paths Jul 24 19:00:21 kergoth: feel free to finish it ;) Jul 24 19:12:37 'er.. Linux rant of the day.. but it may affect us: Jul 24 19:12:38 http://marc.info/?l=linux-kernel&m=140622764926666&w=2 Jul 24 19:12:49 basically he says gcc-4.9.0 is broken when compiling the kernel.. :P Jul 24 19:18:49 yikes Jul 24 19:20:53 yup.. pretty much Jul 24 19:35:19 * zeddii is already watching it, as are a lot of us. it'll get worked out :P Jul 24 19:35:38 * zeddii finishes pushing 3.16 support and gcc 4.9 .. right in the death zone ;) Jul 24 19:55:08 kroon: kergoth yeah Jul 24 20:09:18 hi kergoth Jul 24 22:21:59 halstead: pidge: Around? Jul 24 23:54:37 hi manuel_ Jul 24 23:58:57 hi Jul 25 00:45:14 fray: well thats a typical example of kernel needing more than C standards and more than ABI Jul 25 00:45:50 its not first time it would have happened, I wonder how a compiler other than gcc would get to compile kernel its a tall task **** ENDING LOGGING AT Fri Jul 25 03:00:00 2014